Issuu on Google+

The

E15

Initiative

Strengthening the multilateral trading system

Are Global Value Chains Really Global? Policies to Accelerate Countries' Access to International Production Networks

e15 Expert Group on Global Value Chains


Are Global Value Chains Really Global? Policies to Accelerate Countries’ Access to International Production Networks Antoni Estevadeordal Inter-American Development Bank

Juan Blyde Inter-American Development Bank

Kati Suominen Center for Strategic and International Studies

This version: October 2012 Paper prepared for the ICTSD-IDB E-15 expert’s dialogue on GVCs Abstract The term “global value chain” (GVC) is frequently used to refer to a multi-country process in which different stages of production are carried out in specialized plants in different parts of the world. But casual evidence of this phenomenon suggests that these global linkages might be rather regional, as they seem to concentrate geographically around North America, the European Union and East Asia, with other countries in the world remaining mostly on the sidelines. This paper seeks to document more systematically whether this is the case and if so, why. Using measures of trade in value added, we first provide evidence on the extent to which global value chains are regionally oriented. Then we ask what are the forces shaping the current patterns of trade in GVCs and in particular we examine the role of transport-related factors and trade-policy measures. In light of this evidence, we then assess the policy options for developing countries aiming to improve their participation in crossborder production sharing. JEL No. F10, F23, L23 Key words: international production networks, logistics infrastructure, trade agreements, rules of origin

We would like to thank Kun Li and Julieth Santamaria for excellent research assistance. The views and interpretations in this paper are strictly those of the authors and should not be attributed to the InterAmerican Development Bank, its Board of Directors, or any of its member countries  Correspondence address: Antoni Estevadeordal. Inter-American Development Bank, 1300 New York Ave., NW, Washington DC, 20755, U.S. Phone: (202) 623-2614. Fax: (202) 623-2169. E-mail: antonie@iadb.  Correspondence address: Juan Blyde. Inter-American Development Bank, 1300 New York Ave., NW, Washington DC, 20755, U.S. Phone: (202) 623-3517, Fax (202) 623-2995. E-mail: juanbl@iadb.


Executive Summary The world economy has recently seen an increasing trend in international production fragmentation, or a geographic separation of activities involved in producing a good or a service across two or more countries. The resulting international organization of production has substantially increased interdependencies among economies around the globe and has translated into a fast growing trade in intermediate inputs and services. The process of fragmentation tends to eliminate the need to gain competency in all aspects of the production of a good and allows developing countries to enter into the network of cross-border production sharing by specializing in just one or a few stages of the production activities involved in making a final good. This type of specialization allows developing countries to participate in a finer international division of labor. In its absence, emerging economies would have to master entire production processes in order to become strong competitors in world markets. However, global value chains (GVCs) have not been spread evenly across the world as the main networks concentrate around East Asia, Europe and North America. Many developing countries in Africa and Latin America in particular remain at the sidelines in cross-border production sharing. Indeed, the regional bias in GVC structures does not mean that GVCs as currently structured are optimal: rather, we find that the regional bias stems from high transport and logistics costs that discourage value chains spanning long distances, and by trade policies, first and foremost preferential trading arrangements, which have often been formed among neighboring countries. While being a member of a trade agreement does not necessarily impede a country to develop supply chains with non-member countries, most PTAs have rules of origin that disincentivize the use of materials from outside the bloc. There are a number of multilateral and regional measures to reduce distortions in GVCs and encourage developing nations’ access to global production sharing. First, while important for trade in general, the quality of the logistics systems are particularly relevant for participating in cross-border production processes in which low-inventory cost strategies, such as just-in-time delivery services, continue to be the norm.. The WTO should pay keen attention to several other policy areas that affect transport costs in developing nations, including trade facilitation, Aid for Trade, and trade in transport services. Second, cumulation of production across the various PTAs criss-crossing the world can be quite effective to reduce the distortive effect of rules of origin and spur cross-border production sharing. Granted, multilateral tariff and non-tariff liberalization would also reduce the bite of restrictive rules of origin – and in general encourage trade intermediates. There is also a need for further analytical work on factors that arbitrate GVC structures. For example, more detailed evidence on how specific aspects of the logistics infrastructure (e.g. port and airport efficiency, information and communication infrastructure, custom procedures) interact with the participation of countries in GVC is an important issue to examine further. Another important area is optimal architecture of trade agreements – and balancing the need to curb trade deflection and the importance of keeping rules flexible enough to allow for offshoring opportunities.

2


1. Introduction The world economy has recently seen an increasing trend in international production fragmentation, the geographic separation of activities involved in producing a good or a service across two or more countries. The resulting international organization of production has substantially increased interdependencies among economies around the globe and has translated into a fast growing trade in intermediate inputs and services (Yeats, 2001; Hummels, Ishii and Yi, 2001, UNCTAD, 2004). There has certainly been no shortage of names for this phenomenon, from offshoring, global value chains and international production networks to slicing the value added chain (Krugman, 1995), disintegration of production (Feenstra, 1998), delocalization (Leamer, 1996) or the great unbundling (Baldwin, 2006). All this reflect a rush by economists and business people alike to cope with a fast moving trend that is changing the world’s trade and production patterns. From the point of view of developing countries, the increase in international product fragmentation provides opportunities to engage in international trade transactions that were virtually not available before. The process of fragmentation tends to eliminate the need to gain competency in all aspects of the production of a good and allows developing countries to enter into the network of cross-border production sharing by specializing in just one or a few stages of the production activities involved in making a final good. This type of specialization allows developing countries to participate in a finer international division of labor, and in its absence, emerging economies would have to master entire production processes in order to become strong competitors in world markets. For example, in stark contrast from the harsh development process followed by Germany or Japan in which entire supply chains were built domestically, countries like China or Vietnam are following a different path towards industrialization based on entering existing international supply chains (Baldwin, 2011). But global value chains (GVCs) have not been spread evenly across the world as the main networks concentrate around East Asia, Europe and North America. We find such regional biases owe mostly to trade costs and trade policies. The implication of the regional bias is that many developing countries especially in Africa and Latin America remain at the

3


sidelines in cross-border production sharing, and global value chains are hardly optimized. Since participation in GVCs entails new opportunities for industrialization and development, it is important to. This paper seeks to deepen the understanding of the obstacles to optimal GVC structures and to access of countries to GVCs. The paper is organized as follows. In section 2 we motivate the analysis by summarizing relevant empirical and theoretical literature. In section 3 we present systematic evidence of the extent to which GVCs are regional. An econometric model is developed in section 4 to measure the impact of specific drivers of production location on the regional patters of GVCs. The section also discusses what the findings of this model imply for developing countries seeking to join GVCs. Section 5 discusses multilateral and regional policies for optimizing GVCs. Section 6 concludes. 2. Empirical and theoretical background Most of the evidence regarding global value chains consists on cases involving countries in similar regions. The examples abound from Mexico’s maquiladoras and Canadian suppliers linked to US multinationals (MNCs) to Japanese firms outsourcing production processes in East Asia. Even in the well documented cases of the US-designed iPods and iPhones, most of the actual manufacturing process takes place “regionally� with China assembling parts and components provided by Japan, Korea and Taiwan (Dedrick, Kraemer and Linden, 2008, 2012). The anecdotal evidence is equally supported by more systematic analyses like those using microdata of MNCs. For instance, Kimura and Ando (2005) show that 80% of MNCs in Japan locate foreign affiliates in East Asia and more than half of all the affiliates are located solely in this region. This is not to suggest the absence of truly global supply chains with firms linked across far away regions, but existing evidence tends to support the claim that the majority of international production networks are regionally oriented. While transport costs is the main factor that comes to mind for rationalizing the regional character of many supply chains, casual evidence also suggests that many regional supply chains are intrinsically related to certain agreements and/or arrangements that occur across nearby countries. For instance, before the 1965 US-Canada Auto Agreement, trade in auto

4


parts between these two countries practically did not exist. After the 1965 agreement reduced the tariffs to zero, auto trade soared igniting a successful US-Canada auto supply chain in which 60% of US auto exports to Canada are engines and parts, while 75% of Canada auto exports to the US are finished cars and trucks (Hummels, Rapoport and Yi, 1998). The emergence of “factory Asia� is another example in which the surge of Asian supply chains coincides with a series of actions that even though were not always well-organized policies designed for the formation of international production networks, they were adopted regionally and may have helped spurred the growth of regional cross-border production sharing in this part of the world. For example, starting in the 1970s many countries coincide in implementing policies for the aggressive attraction of FDI –the backbone of many supply chains- and after the Asian currency crisis in 1997-98, the trading bloc became much more integrated (Kimura, 2006) deepening trade linkages within the region particularly in parts and components. Today, for instance, low protection levels on semi-processed products are a common factor across the Asian region (WTO IDE-Jetro, 2011). The enlargement of the EU provides yet additional evidence on the relationship between common regional actions and the formation of regional production networks. For instance, using trade data, Curran and Zignago (2012) show that the reliance of the EU industry on other member states -which has been traditionally large- has only increased since the enlargement. The process of enlargement, which required the new member states to apply the same common external tariff of the bloc, triggered unprecedented flows of FDI from the old to the new members with the main recipients being Hungary, the Czech Republic and Poland, the three countries that since then have become crucial parts of the European supply chain (Karkkainen, 2008). In the next section we provide systematic evidence on the extent to which global value chains are regionally oriented using world data on trade in value added. Before moving to this section, however, it is useful to review the relevant literature behind the formation of

5


global value chains in order to highlight some of the forces that may explain regional patterns behind the international organization of production. During the last two decades the literature on the so call theory of fragmentation or offshoring has been growing rapidly. Following the work in Jones and Kierzkowski (1990), economists have been writing models that explicitly recognize the fact that firms are increasingly fragmenting production processes in various stages or tasks and moving them to places with different location advantages (Jones and Findlay, 2000, 2001; Jones & Kierkowski, 1998, 2000, 2001; Deardorff, 2001a, b; Grossman and Rossi-Hansberg, 2008). These studies examine the main forces behind the international organization of production. Most of the models in this literature share features from an earlier literature on FDI,1 namely that firms will fragment production across different countries to arbitrage international differences in factor prices (Helpman, 1984 and Helpman and Krugman, 1985).2 The basic rationale behind the theory of fragmentation is as follows: in traditional production processes, inputs are organized and combined to generate final outputs in the same location. In the presence of many inputs, coordination is normally necessary and proximity helps keep the costs of coordination low. But if firms could separate the production process into various production blocks and relocate them in places with lower factor costs, the total costs of production could be lowered. Thus, firms may unbundle their production processes as long as the saved production costs arising from the fragmentation process compensate the additional costs of coordinating remotely located production blocks plus the costs of moving these production blocks around.

1

The new models of fragmentation are generally not limited to examine multinationals exclusively. The main predictions of these models tend to apply to companies that fragment production internationally regardless of whether this is done within the boundaries of the firm or through independent suppliers. A more recent strand of the literature examines the more specific issue of whether the fragmentation of production occurs within the boundaries of the firm or through an independent supplier (Antras, 2003; Antras and Helpman, 2004, 2008). This is called the internalization decision. 2 This class of models is called the vertical model of FDI and it was developed in parallel to the horizontal model of FDI in which the motive behind the MNC is to save on trade costs associated with exporting by setting up foreign subsidiaries producing similar goods to those produced at home (Markusen, 1984 and Horstmann and Markusen, 1987). Later on, the knowledge-capital model was developed allowing for a simultaneous horizontal and vertical motives of FDI (Markusen, 1997)

6


This framework highlights the main forces behind the international unbundling of production, namely comparative advantage considerations like differences in factor prices across countries, as well as the overall costs of coordinating activities and moving the various inputs between the supplier’s country and the parent country. In this respect, trade impediments like tariffs as well as the costs of transportation are likely to be major factors behind the costs of moving the inputs across borders. It is then reasonable to expect that countries that are part of a regional trade agreement are more likely to engage in cross-border production sharing, particularly because of the proportionally larger effects of lowering a tariff rate to a production process that cross borders many times -as it is often the case in international supply chains- as oppose to a final good that cross borders only once.3 Additionally, due to the impact of transport costs, large distances can potential erode the saved production costs that emanate from the fragmentation and relocation of production limiting the range of countries that could join a given supply chain. An additional factor that may limit the physical distance between buyers and suppliers is related to uncertainty. Uncertainty in the delivery of any single component of a supply chain can have quite disrupted impacts in the production of a final good as entire production lines might be shut down until all the necessary inputs have arrived. Companies can face this uncertainty by holding large quantities of inventory but modern supply chain practices are increasingly moving towards low inventory-holdings in an effort to cut costs, part of the so call lean production strategies. To the extent that uncertainty in delivery increases with distance, clustering would be the most likely outcome in order to assure timely delivery of all the components. This is indeed the prediction of a recent theoretical analysis by Harrigan and Venables (2006). Based on the above discussion, and after presenting evidence on the regional patterns of GVC participation in the next section, section 4 shows an econometric model that examines

3

This notion is formally developed by Ishii and Yi (1997). They show that tariff reductions have a proportionately greater effect on vertical trade involving goods produced sequentially in multiple countries relative to goods produced entirely in one country.

7


the relationship between regional patterns of trade in GVCs and specific drivers of GVC location. 3. The Patterns of Trade in Value Added One way to account for the participation of countries in global supply chains is to trace the value added of each source country in a globally integrated production network. Studies like this have emerged on a case by case basis, from the iPod and iPhone (Dedrick, Kraemer and Linden, 2008, 2012) to the less technologically intensive but still widespread multi-country production of a Barbie doll (Tempest, 1996). The information in these case studies is very rich as they show which countries participate in the supply chain of a particular good and how much value they add to its production. The studies have revealed, for example, that even though China exports the iPod and accrue in its trade statistics the full value of this product, the country only contributes to 3.8% of the value as many other countries also participate in the production. This case by case examination of specific international supply chains is very revealing but the approach is so data-demanding that it would be unfeasible to cover the supply chain of every good in which a country co-participates in its production. This makes the technique impractical to generate aggregate measures of the participation of countries in GVCs. Recently, however, a new literature has emerged combining input-output tables with bilateral trade statistics in order to trace the value added of a country’s trade flows (e.g. Hummels, Ishii and Yi, 2001; Johnson and Noguera, 2012a, 2012b; Miroudot and Ragousssis, 2009; Koopman, Wang and Wei, 2008; Koopman, Powers, Wang and Wei, 2010; De La Cruz, Koopman and Wang, 2011). The literature has evolved rapidly and has produced an array of indicators assisting in the quantification of the extent to which countries participate in cross-border production sharing. One of the early indicators of participation in GVCs coming from this literature is based on the notion of vertical specialization. In essence, vertical specialization refers to the use of imported inputs to produce goods that are later exported, a notion that captures the idea that various countries are linked sequentially to produce a final good (see Hummels, Ishii

8


and Yi (2001). More recently, the concept of foreign value-added in exports has been introduced which is a refined measure of vertical specialization in which the emphasis is on the value-added from other countries employed in a country’s exports (Koopman, Powers, Wang and Wei, 2010). Foreign value-added of exports is nowadays a common measure to examine for the participation of a country in GVCs. This is one of the indicators that we employ. The appendix at the end of this paper presents the precise definition of the concepts employed as well as the data sources used to construct the indicators. In what follows we will examine first the overall GVC participation of major regions in the world (Europe, Asia-Pacific, North America and Latin America) and second, we will analyze the regional bias of this participation, that is, the extent to which the GVC participation takes place among countries of the same region versus countries from other regions. The appendix also shows in detail the list of countries included in each of the regions which also depends on the availability of the data employed. Nevertheless, it is worth mentioning that Europe comprised the 27 members of the European Union. Asia-Pacific combines the ASEAN countries with the East Asian countries and also includes Australia and New Zealand. Finally, the separation between North America and Latin America is obvious except for Mexico. While Mexico is a Latin American country, it is much more integrated in supply chains with the US and Canada than with Latin America and thus it is more natural to include Mexico in the North America group. This is what we do in this paper.4 Overall participation in GVCs In figure 1 we show the foreign value-added of exports for each region. More precisely this is the simple average of the foreign value-added of the exports of all the countries in the particular region. It can be seen that on average countries in Europe exhibit the largest foreign value-added followed by Asia-Pacific, North America and lastly Latin America. One aspect that might seem puzzling at first is the relatively small value of foreign valueadded in North America relative to the values observed in Europe or Asia. This, however, can be explained by two factors. On the one hand, large countries like the US or Canada, 4

It is worth noting that including Mexico in the Latin America group does not change qualitatively any of the results

9


usually find more factors and resources within their own borders than small countries. It is well known that this fact generates small values in traditional trade measures like “openness” and in the same way it generates small values on our measure of foreign valueadded. The second factor is that in some industrialized countries, primarily in the US, some of the value added incorporated in their imported inputs originates from that country itself. Consider the example of a US firm that exports to Mexico materials for processing which are then re-exported back to the US and used as intermediate inputs in a final good that is later exported to other countries. The original value-added from the US that is incorporated in these imported inputs from Mexico is not included in our calculation of foreign value-added. This indeed is called in the literature “reflected domestic value-added” or the domestic value-added that returns home (Koopman, Powers, Wang and Wei, 2010). In countries, like the US, in which the domestic value-added that returns home is large, the foreign valueadded tends to be relatively smaller than in other countries. But this does not mean that the country does not participate in cross-border production sharing, it only means that the participation of the US in GVCs is much more complex than the participation of other countries, as the US enters in the supply chain at various points of the chain. In Figure 2 we add this “reflected value added” to the measure of foreign value added presented before, and we can see that the participation of the North America region indeed increases substantially. Another measure that is useful in assessing how countries participate in supply chains is the GVC position (see the appendix). Koopman, et al. (2010) measure GVC position as the ratio of indirect value-added to foreign value-added where indirect value added is the value added of a country that is embodied in the exports of its partners. We employ the same measure here When the GVC position is high, the country tends to participate more as a provider of value-added than as a recipient of foreign value-added; therefore, the country is relatively upstream in the chain. Conversely, when this measure is small, the country tends to participate more as a recipient of value-added than as provider and thus the country is relatively downstream in the chain. Figure 3 shows the average GVC position by regions. The Latin American region has the largest GVC position which is not surprising. Latin American countries tend to export natural-resource intensive goods and these types of goods tend to

10


be located upstream in the chain. Europe, at the other, has the smallest value reflecting a position in GVCs much closer to the end of production process. North America and AsiaPacific are in the middle of the range indicating a mix of production process in which the participations in GVCs as providers and as recipients of value-added is somewhat balanced.

Regional patterns of trade in GVCs The previous indicators were useful in highlighting differences across regions in terms of their overall participations and positions in GVCs. However, the indicators did not say anything about the regional bias of the involvement in GVCs. For instance, we do not know if the foreign value-added of the Asian exports come mostly from countries in Asia or from other regions. The usefulness of the methodology employed in this paper based on international input-output tables is that we can precisely track the origin of the foreign value-added by country/region source. This allows us to move beyond the casual evidence mentioned in section 2 and present more systematic evidence of the extent to which participation in GVCs is regionally biased. This is what we do now. Figure 4 shows for each region the contribution to its foreign value-added made by the countries in its own region and by the countries from other regions. The results show unequivocally a large regional biased in GVC participation. On average, about half of the foreign value added originates from countries in the same region. In Europe, Asia-Pacific and North America, for example, the within-regional contribution to foreign value added is 51%, 47% and 43%, respectively, and in none of these cases the contribution from extra-regional sources reaches 20%. Only in Latin America the contribution of another region, in this case North America (28%), is practically similar to the contribution of its own region (27%) but even in this case the contributions of more distant regions like Europe and Asia are much smaller. These results point to a strong role for proximity in explaining the patterns by which countries unbundle their production processes and locate them abroad. As mentioned in section 2, there can be various mechanisms that could explain why proximity may matter,

11


most notably the costs of shipping bundles of goods across space and the existence of regional trade agreements among nearby countries. In the next section we present an econometric model to disentangle the role of each of these forces while controlling for other elements as well. 4. Determinants behind regional patterns of GVCs In this section we estimate an empirical model to isolate the impact of specific drivers behind the regional patterns of GVCs. The model is based on a gravity equation, the workhorse of empirical international trade. Gravity equations, which have been shown to have theoretical underpinnings5, are typically estimated using gross trade flows. More recently, however, they have also been employed to examine trade in value added (Johnson and Noguera, 2012b). The specific gravity model that we employ takes the following form: (1) where

is the foreign value added from country j in the exports of country i;

fixed effects for country i and country j, respectively, and

and

are

is a vector of bilateral variables,

is a variable that captures preferential trade agreements. The formulation

follows others in using individual country fixed effects to estimate trade equations (Feenstra, 2004; Eaton and Kortum, 2001, 2002).6 More specifically, the vector comprises a series of variables that are standard in gravity models. These variables are the bilateral distance between both countries and dummy variables for same border, same language and same colonial ties. The dummy variable

is equal to one if the countries

are part of the same preferential trade agreement and zero otherwise. In this paper we are particularly interested in the coefficients for distance and for the trade agreement variables. Note that in our specification the fixed effects

and

will control

for any country characteristics like size (e.g. GDP, population, area) or level of development

5

See Anderson and van Wincoop (2003) and Eaton and Kortum (2002) These individual country fixed effects play the same role as the multilateral resistance index introduced by Anderson and van Wincoop (2003). Additionally, potential econometric problems related to exogeneity and omitted variables are largely reduced by using these fixed effects (Anderson and Yotov, 2012) 6

12


(e.g. GDP per capita), that may also affect the extent to which countries use foreign valueadded. The results of our baseline estimation are shown in Table 1. As noted in columns (1) and (2), countries that share the same border, language and colonial ties are more likely to engage in cross-border production sharing. More importantly, the results show that physical distance substantially decreases foreign value-added. In column (2), for example, the results imply that an increase of 10% in distance reduces foreign value-added, on average, by around 67%. Physical distance is a proxy for the costs of transportation and thus the finding support the notion mentioned in section 2 that if these costs are large they can erode the saved costs from locating blocks of production in other countries, particularly if they are far away. Another finding from column (2) is that even after controlling for the effect of distance, foreign value-added is positively and significantly related to PTAs. Membership in the same PTA increases foreign value-added, on average, by around 15%. In other words, countries will source 15% more of their foreign value-added from members of the same PTA than from non-members. Therefore, the fact that trading across borders within the same PTA does not add extra-duties creates an incentive to source part of the production process from countries that share the same trade agreement. It is worth mentioning that while the role of distance is clear in explaining the regional bias of GVC participation, the role of PTAs is less obvious; after all, PTAs have been signed between countries as distant as the US and Bahrain. More often than not, however, PTAs are signed between nearby countries or among countries in similar regions; therefore, sharing a trade agreement often represents an incentive to source materials from proximate countries in addition to the incentives offered by transportation costs savings.

Discussion and further evidence The econometric evidence shows that proximity matters. Global supply chains are less likely to flourish the larger the distance among participating countries. This strong result opens up 13


some questions regarding the options for many developing countries that are not close to major GVC regions but that are aiming to improve their participation in them, like countries in LAC. At least three issues are worth mentioning regarding this. First, a remote country/region is not entirely forbidden to join an international supply chain because of the long distance. However, the results imply that some form of compensating for the high costs of transportation would probably be required. Most likely this compensation will take place with savings in production costs arising from strong comparative advantages. Therefore, for many developing countries, like in LAC, accessing GVCs in other regions is likely to take place, at least initially, in sectors in which the comparative advantages are the strongest. Second, transport costs are not solely determined by distance. The costs of transportation also depend on issues like volume, the level of containerization of the cargo, the degree of competition among shipping companies, and the quality of the transport-related infrastructure, among others. It has been shown, for example, that around 30% of the larger freights rates of LAC’s exports to the US relative to those from Europe can be explained by differences in port efficiency (see Mesquita Moreira, Volpe and Blyde, 2008). What the results then imply is that for far away countries to join GVCs, issues like improving port or airport efficiency are likely to be more important -in order to offset the impact of distancethan for nearby countries. Finally, countries away from major GVC regions can seek to develop their own regional value chains. This could be the case, for instance, across countries in LAC in which some regional value chains already exist. But even if these were the type of supply chains that were more ripped to emerge in LAC, it is still important to be aware of the differences that exist in terms of distances across countries in LAC versus the distance across countries in other regions. For example, the average bilateral distance across all the East Asian + ASEAN countries is around 2,400 Km while the average distance across the countries in LAC is 3,000 Km, and if we include the US and Canada, the distance across the Americas is 3,200 Km, or 30% more than in Asia.7 Therefore, even for the prospects of intra-regional supply chains in the Americas, the vast distances of the continent impose a challenge. Additionally, because of the geography, most of the shipments across countries in Asia take place through the ocean while within LAC, many shipments take place 7

We use great-circle distances between the capitals of the countries.

14


by land transportation, a mode with lower economies of scale relative to the maritime mode and thus with higher per unit freight costs. True, land transportation is the main shipping mode within Europe but the distances in this part of the world are much smaller with an average bilateral distance among all the 27 European countries of 1,400 Km which is less than half of the average distance across the Americas. Again, the issue of long distances does not necessarily mean that the prospects for the creation of regional supply chains are doomed to fail. It means that everything about connectivity, including the efficiency of port and airport infrastructure as well as the quality of internal road networks is relatively more important for LAC than for countries in other regions that are closer to each other. The role of trade agreements is another issue emerging from the econometric results that is worth discussing further. As mentioned before, the fact that many trade agreement occurs between countries in similar regions imply that sharing a trade agreement adds an additional layer of incentives, beyond transport costs considerations, to co-locate production in nearby countries. So what does this implies for the prospects of countries that are not members of the PTAs? In principle, being a member of a PTA does not necessarily impede a country to develop supply chains with non-member countries particularly if tariff rates between them are not very large. However, most PTAs have rules, in the form of rules of origin (RoO) that would disincentivize the use of materials from outside the bloc particularly if they are employed to produce final goods that are later exported to other members of the PTA. In some cases, however, the RoO are flexible enough for countries to engage in cross-border production sharing even with members outside the PTA, therefore, it is worth examining the issue of RoO in more detail. RoO are critical parts of PTAs because they establish the conditions that a product must satisfy to be deemed as originating from the country seeking preferential access (Estevadeordal and Suominen, 2006). They are primarily used to prevent trade deflection – that is, to avoid products from non-participating countries reaching a high-tariff PTA partner via the transshipment of the product through a low-tariff PTA member. But depending on how they are designed, RoO can have quite important implications in the way firms choose the location in which they fragment production.

15


The most obvious implication, as mentioned before, is that RoO restrict outsourcing from countries outside the PTA. Take, for instance, a firm producing a good entirely in country A and selling it tariff free to country B within the same PTA. If the firm now decides to use inputs from a country C outside the PTA, then depending on the RoO of the agreement, the same final good that it is exported from A to B will now pay import duties. Therefore, the RoO can disincentivize outsourcing from outside the PTA and in particular, if the final destination of the good is within the PTA. RoO can also limit the outsourcing of production processes among countries that have parallel or overlapping PTAs. Consider the example of a country A that has separate PTAs with two countries B and C. Any final good produced in B or C would have tariff free access to country A; however, a good produced in B using inputs from C and exported to A would not. In this case, the RoO create disincentives for firms in country B to outsource part of the production process to country C even if it makes economic sense to do so. Finally, RoO can even limit the outsourcing possibilities among countries of the same PTA. Take for instance, the case of NAFTA. In this agreement the relevant RoO implies that a final good produced in the US with inputs from Canada can be exported to Canada with zero import duties in the same way as if the good was entirely produced in the US. However, if the inputs used in the US come from Mexico –another NAFTA country- the final good exported to Canada will have to pay an import duty. Again, this creates disincentive to relocate part of the production process from the US to Mexico. While RoO can impose limits to the range of countries that can participate in the international fragmentation of production, it is also possible to reduce these limits through various instruments, for instance, with flexible cumulation rules (such as diagonal cumulation), with higher de minimis levels, or by allowing for duty drawback. Cumulation in general means that inputs from trading partners can be used in the production of a final good without undermining the origin of the product. In the specific case of diagonal cumulation, inputs from say a non-member country can be used for exports geared towards members of the agreement without paying extra-duties. De Minimis rules allow a specified percentage of non-originating products to be used in the production process without

16


affecting the origin status of the final product. A duty drawback can be used to return the payment of duties applicable to the non-originating material employed in the production of a final product that is subsequently exported to other members of the agreement. It is important to note that because RoOs can restrict the fragmentation of production even within the same PTA, depending on how they are crafted the mechanisms mentioned above (cumulation, de minimis, duty drawbacks) can bring incentives to encourage more crossborder production sharing among the countries of the same PTA and/or between them and non-PTA countries. It is worth noting, however, that the capacity of these mechanisms to incentivize trade in general and in particular trade in inputs is not an issue that has being analyzed empirically before. From the results in Table 1, for example, we know that a country is more likely to obtain foreign value-added from a country that shares its same PTA, but we do not know what extra incentives mechanism like diagonal cumulation can create to foster the participation of a country in GVCs. This is an issue that we explore now. To gain some intuition before we proceed with the econometric analysis, let us consider the following example. China is a country that has 10 PTAs, including one with Chile. None of the PTAs that China has with countries different from Chile allows diagonal cumulation with Chile. This means that an export from China to any of its PTAs will enter with zero tariff if the good is produced entirely in China but will pay a duty if it uses materials from Chile. This situation implies the following: on the one hand, the agreement with Chile means that when China sources imported inputs to be employed for domestic final demand or for exports to countries that are not part of China’s PTAs, China has more incentives to trade with Chile than to say Bolivia, a country with whom China does not have a trade agreement. This effect is captured by the PTA dummy that we include in the gravity model in equation (1) and it is also the typically bilateral effect that is analyzed in the empirical literature of free trade agreements and trade. However, when it comes to importing goods that will be subsequently used in the exports of China to members of its PTAs, China does not have an incentive to import from Chile more than from Bolivia even though it has an agreement with Chile. This is because none of the PTAs of China allow cumulation from Chile or by the same

17


token from Bolivia. In other words, China’s exports to other members of its PTAs will need to pay the same extra duties for using non-originating inputs from Chile or from Bolivia. Therefore, beyond the incentives created by the preferences that China grants to Chile with its own bilateral PTA, China does not have any additional incentives to source inputs from Chile relative to Bolivia. Now, let us complete the example by introducing another country, Thailand. Thailand is a member of the ASEAN agreement with whom China has a PTA. It is also the case that diagonal cumulation is allowed between the ASEAN countries and China. This situation implies the following: first, when China sources imported inputs to be employed for domestic final demand or for exports to countries that are not part of China’s PTAs, China has more incentives to trade with Thailand than to trade with Bolivia because it has a trade agreement with the first country and not with the latter. This is similar to the incentives that China has to trade with Chile more than with Bolivia due to the direct effect of having a bilateral trade agreement. Again, this is the bilateral impact captured by the PTA dummy in the gravity equation. The second effect now implies that when it comes to importing goods that will be subsequently used in the exports of China to other members of its PTAs, China has an additional incentive to import from Thailand more than from Chile or from Bolivia because it can cumulate materials from Thailand to export to other ASEAN countries while it cannot do so from Chile or from Bolivia. Therefore, the diagonal cumulation gives an extra incentive to import from Thailand. Note that this incentive can take place regardless of whether there is a free trade agreement between Thailand and China. In what follows, we present an empirical exercise that examines the role of diagonal cumulation on cross-border production sharing. The exercise consists on augmenting the gravity model in equation (1) as follows: (2) where

is equal to 1 if the importing country i has a trade agreement with third

countries that permit cumulation with country j and zero otherwise, and the rest of the variables are the same as before. Note that the traditional bilateral effect of having a trade agreement will still be captured by the PTA variable. In other words, the PTA variable will

18


still measure the incentives for country i to source inputs from country j because they share a trade agreement. But in addition to this effect, now the CUM variable will capture the additional impact that involves a relationship between three parties. That is, CUM will measure the incentives for country i to source inputs from country j because country i can cumulate these inputs to export goods to third countries. Note that this modeling can encompass many different situations. For instance, country i and j might have a PTA, and this PTA might be the same as the PTA that country i have with third countries. It can also be the case that the PTA between countries i and j is different than the PTA that country i have with third countries. Finally, it is also possible that country i and j does not share a PTA. In all these situations, CUM will be equal to 1 if country i has a PTA with third countries that permit cumulation from country j. Constructing the cumulation dummy is not an easy task. It involves, for each country i, analyzing each of its PTAs and identifying for each agreement whether it permits cumulation with each of the countries in the sample. Accordingly, we restrict this exercise to a smaller set of countries. In particular we examine the country members of the ASEAN FTA (AFTA)8, the ASEAN-China (ACFTA)9, the ASEAN-Korea (AKFTA)10 and the Trans-Pacific Partnership.11 More specifically, the sample of country i will be restricted to the countries in these agreements while the sample of country j includes all the countries as before. Note that the countries in each of these agreements are allowed to engage in diagonal cumulation among themselves. Note also that because within each agreement, any pair of countries share the same PTA and are also allowed to cumulate with other members of the agreement, the variables PTA and CUM will be equal to 1 among all the country pairs of each of these agreements. The identification of the effect on cumulation comes from the fact that many of these countries also have PTAs with other nations with whom they cannot cumulate. In those cases the PTA dummy is equal to 1 while the CUM dummy is equal to zero.12

8

Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand and Vietnam ASEAN countries + China 10 ASEAN countries + Korea 11 Brunei, New Zealand, Chile and Singapore 12 It is also the case that the countries in these agreements do not have PTAs with many other countries in the sample 9

19


The results of this exercise are shown in Table 2. Columns (1) and (2) depict similar specifications as those in columns (1) and (2) of Table 1; that is, without including the variable CUM. Note from these columns that again the PTA dummy has a positive and significant impact on foreign value-added with a coefficient that is comparable to the results in Table 1. In column (3) we add the variable to capture the effect of cumulation. The results in this column indicate that the PTA variable is still positive and significant with a coefficient that decreases only slightly relative to column (2). Most importantly, there is this additional impact from the cumulation mechanism that is also positive and significant. The results imply that countries will source 9% more of their foreign value-added from members of the same PTA and an additional 20% if cumulation with that country is possible. Column (4) presents an alternative specification to examine the same issue. Specifically, we include a PTA dummy that is equal to 1 if country i has a PTA with country j but no cumulation is possible (PTA - No cumulation) and another PTA dummy that is equal to 1 if country i has a PTA with country j but cumulation is possible (PTA - Cumulation). In essence this specification presents more clearly the differential impact of PTAs that allow cumulation from PTAs that do not allow cumulation. According to the results countries will source 9% more of their foreign value-added from members of the same agreement if the PTA does not allow for cumulation while they will source 30% more from members of the same agreement if the PTA does allow for cumulation. This result is equivalent to the one in column (3). Both results support the notion that the capacity to cumulate adds extraincentives to trade and in particular to engage in cross-border production sharing. Obviously, because in this exercise the diagonal cumulation mechanism only takes places among the countries of the same PTA, the findings only provide support to the notion that diagonal cumulation significantly encourages cross-border production sharing across PTA countries. But in general terms the results suggest that diagonal cumulation can be an important mechanism to promote international linkages between PTA-members and nonPTA-members provided that such diagonal cumulation is allowed. The general idea is that even though a country is not part of a PTA, allowing cumulation with that country can go a long way to foster its insertion into regional supply chains. This is then one mechanism by

20


which developing countries that are not members of certain PTAs could benefit as their likelihood to access international/regional production networks improves. 5. Implications for the WTO GVCs are mostly regional. This is due primarily to transportation costs that rise with distance, and trade agreements, which have first and foremost been forged among neighboring countries. This regional bias in GVCs has two implications. First, some countries (typically developing countries in Latin America and Africa) have remained at the periphery of GVCs: they are distant from the main GVC clusters and do not necessarily share trade agreements with the “GVC hub regions.â€? Second, some the regional clustering of GVCs is due to geographic factors that cannot be altered, but much is due to trade policy and transportation costs that can. Granted, these variables will likely evolve with changes in energy costs (which, if increasing, can incentivize shorter value chains), education and information technology (which, if improving, can obliterate distances), and new trade agreements (which are increasingly struck across continents). However, an optimal scenario where GVCs are unconstrained by trade or transportation barriers has yet to be reached. Only a multilateral approach can effectively encourage globalization of value chains: there is unique value in a multilateral approach that includes all regions, and in a comprehensive approach across trade disciplines. This is something only the WTO can accomplish. There are at least six major areas of work to reduce transportation costs and trade barriers in order to optimize GVCs: 

Trade facilitation: Trade facilitation is key to fluid trade in intermediate products and particularly critical for developing nations, which stand to gain considerably from policies and measures that help fuel trade. Trade facilitation is also a central issue in the Doha Round. However, the negotiations have been quite narrow in scope and the implementation of their outcome will take a long time. In light of the urgency of trade facilitation to the WTO members and developing nations in particular, the WTO members should agree to implement the Trade Facilitation Agreement, 21


possibly as a plurilateral agreement not requiring a formal conclusion of the Single Undertaking. 

Aid for Trade: To a large extent purported to catalyze investments in infrastructure improvements, the global Aid for Trade agenda, an integral part of the Doha Round, is an important complement to trade facilitation agreements in helping developing countries access global value chains.

Barriers to intermediate imports: Protectionist reactions crept up in the wake of the Great Recession, and some nations have supported tariff and non-tariff measures, such as “buy local” provisions, for replacing input imports with domestically made goods, generally in order to encourage production and job creation at home. Policies to keep imports out and production in are, however, ultimately self-defeating, curbing access to the most efficient intermediate goods for production of exports, undermining opportunities to absorb foreign technologies, and eroding participation in GVCs – which would help create jobs. Non-tariff barriers such as regulations and standards can significantly limit trade in intermediates and must be dealt with rigorously at the multilateral level.

Services liberalization: Quality of transportation services affects trade costs and is a major determinant of how GVCs are structured. Transport, communications, and distribution are key services sectors and closely linked to trade costs. Yet barriers to trade in services remain very steep particularly in developing nations. Liberalization of transportation services could significantly enhance these countries’ access to GVCs. Granted, PTAs have progressed well beyond the WTO and GATS in services liberalization; for the WTO to catch up, the members might be well-advised to seek a plurilateral agreement among a substantial coalition of the willing.

Cumulation across PTAs: PTAs reduce the trade policy barriers that otherwise might discourage value chains among the parties; however, PTAs’ rules of origin can undermine the members’ incentives and ability to participate in GVCs. Such problems can be reduced through further multilateral tariff liberalization (which reduces the attraction of PTA preferences and thus need to comply with RoO) as well as cumulation across several RTAs into larger trade areas, such as took place in

22


Europe in the late 1990 and as is occurring in the Asia-Pacific through the TransPacific Partnership negotiations and also in Latin America in the context of the Pacific Alliance. 

Intellectual property: Outsourcing and “slicing” the production chain requires the sharing of a wide array of proprietary knowledge. Inadequate intellectual property protections erode innovators’ profit margins and businesses’ incentives to extend value chains to third countries for fear of IP theft. Concerns about intellectual property have accentuated with recent efforts to incentivize technology transfer to certain nations through compulsory licensing of patents and “indigenous innovation” policies. China for example has established a preference in government procurement to products whose IP is owned and originally trademarked in China – which essentially means exchanging market access to foreign firms for technology transfer to China. This disadvantages foreign property rights holders across the world and discourages production sharing. The immediate priority at the multilateral level is to ensure effective enforcement of the Trade-related aspects of intellectual property rights (TRIPS) agreement.

23


6. Conclusion Rapid technical progress allowing the physical fragmentation of production in various bundles, combined with a general decline in transport costs and an improved capacity of communication and information systems have allowed firms to separate production processes to take advantage of differences in relative prices across the world. While the importance of proximity has declined, it has not disappeared completely: international production networks continue prevalent across nearby countries or in similar regions. The regional bias in the formation of GVCs imposes a challenge for developing countries aiming to participate in GVCs but that are away from these industrial clusters. This paper shows that high transport costs accrued by long distances are a major determinant of countries’ ability to participate in GVCs. This is no reason for fatalism: transport costs are not only about distance, but also about the quality of transport-related infrastructure which include aspects like the capacity of a port to move in and out merchandise without uncertainty, delays and/or damages. While important for trade in general, the quality of the logistics systems are particularly relevant for participating in cross-border production processes in which low-inventory cost strategies, such as just-in-time delivery services, continue to be the norm. More detailed evidence on how specific aspects of the logistics infrastructure (e.g. port and airport efficiency, information and communication infrastructure, custom procedures) interacts with the participation of countries in GVC is an important issue to examine further. The WTO should pay keen attention to several other policy areas that affect transport costs, including trade facilitation, Aid for Trade, and trade in transport services. We also show in this paper that the regional bias in GVCs stems from trade and other agreements among neighboring countries. While being a member of a trade agreement does not necessarily impede a country to develop supply chains with non-member countries, most PTAs have rules of origin that disincentivize the use of materials from outside the bloc. This paper shows that RoOs have quite important implications in the way firms choose the location in which they fragment production, typically restricting outsourcing even from countries of the same PTA. Specifically, we provide some preliminary

24


evidence showing that instruments like diagonal cumulation can be quite effective to reduce the strictness of these rules and spur cross-border production sharing among PTA members. Granted, multilateral tariff and non-tariff liberalization would also reduce the bite of restrictive rules of origin – and in general encourage trade intermediates. Additional analysis on this area should be welcomed to further illuminate on the extent to which this and other mechanisms can also encourage production linkages between members and non-members of a PTA. More generally, the results of this paper suggest that another important area for further discussions on the optimal architecture of trade agreements and the need to balance the trade-off between curbing trade deflection and keeping rules flexible enough to allow for potential offshoring opportunities.

25


References Antras, P., 2003, “Firms, Contracts and Trade Structure” Quarterly Journal of Economics, 118 Antras, P., and E. Helpman, 2004, “Global Sourcing” Journal of Political Economy, 112, pp 552-80 Antras, P. and E. Helpman, 2008, “Contractual Frictions and Global Sourcing” In Helpman et al. 2008, pp 9-54 Anderson, J.E., and Y.V. Yotov, 2012, "Gold Standard Gravity," NBER Working Papers 17835, National Bureau of Economic Research Anderson, James E., and Eric van Wincoop. 2003, “Gravity with Gravitas: A Solution to the Border Puzzle” American Economic Review, 93(1) Baldwin, Richard, 2006, “Globalization: the great unbundling(s)”, report prepared for the Finnish Prime Minister’s Office (in context of EU Presidency 2006). Baldwin, R., 2011, "Trade And Industrialisation After Globalisation’s 2nd Unbundling: How Building And Joining A Supply Chain Are Different And Why It Matters," NBER Working Papers 17716, National Bureau of Economic Research Curran, Louise, and Soledad Zignago (2012) EU enlargement and the evolution of European production networks. Research in International Business and Finance. Volume 26, Issue 2, May 2012 De La Cruz, Justino, Robert Koopman and Zhi Wang, 2011, “Estimating Foreign Value-added in Mexico’s Manufacturing Exports” US International Trade Commission Working Paper No. 2011-04A Deardorff, A.V., 2001a, "Fragmentation in simple trade models," The North American Journal of Economics and Finance, Elsevier, vol. 12(2), pages 121-137, July. Deardorff, A.V., 2001b, “Fragmentation across Cones,” in SvenW. Arndt and Henryk Kierzkowski, eds., Fragmentation: New Production Patterns in the World Economy, Oxford University Press, 2001, pp. 35-51. Dedrick, Jason, Kenneth L. Kraemer and Greg Linden, 2008 “Who Profits from Innovation in Global Value Chains? A Study of the iPod and notebook PCs”, UC Irvine. Paper prepared for the Sloan Industry Studies Annual Conference Dedrick, Jason, Kenneth L. Kraemer and Greg Linden, 2012, “iPhone profits” in Communications of the ACM. http://pcic.merage.uci.edu/papers/2011/Value_iPad_iPhone.pdf

26


Eaton, J., and S. Kortum, 2001, “Trade in Capital Goods,” European Economic Review 45, 1195–1235. Eaton, Jonathan and Samuel Kortum. 2002, “Technology, Geography and Trade” Econometrica, 70(5) Estevadeordal, Antoni, and Kati Suominen, 2006 “Mapping and Measuring Rules of Origin around the World” in The Origin of Goods; Rules of Origin in Regional Trade Agreements, edited by O. Cadot, A. Estevadeordal, A Suwa-Eisenmann, and T. Verdier. Oxford: Oxford University Press Feenstra, R, 1998, “Integration of trade and disintegration of production in the global economy” Journal of Economic Perspective, Volume 2, Number 4. Feenstra, R, 2004, Advanced International Trade: Theory and Evidence, Princeton, NJ: Princeton University Press. Findlay, Ronald and Ronald Jones, 2000, “Factor Bias and Technical Progress”, Economics Letters 68, pp 303-308 Findlay, Ronald and Ronald Jones, 2001, “Input Trade and the Location of Production”, American Economic Review, vol 91 (2), pp 29-33. Grossman, GM, and E. Rossi-Hansberg, 2008. “Trading Tasks: A Simple Theory of Offshoring” American Economic Review 98, pp 1978-97 Harrigan, James and Anthony J. Venables, 2006. “Timeliness and Agglomeration” Journal of Urban Economics 59. Helpman, E., 1984, "A Simple Theory of Trade with Multinational Corporations," Journal of Political Economy 92, 451-471. Helpman, E. and P. Krugman, 1985, Market Structure and Foreign Trade, Cambridge: MIT Press. Horstmann, I.J. and J. R. Markusen, 1987, "Strategic Investments and the Development of Multinationals," International Economic Review 28, 109-121. Hummels, David, Jun Ishii and Kei-Mu Yi, 2001, “The Nature and Growth of Vertical Specialization in World Trade” Journal of International Economics, 54:1; Hummels, D., D. Rapoport and K-M., Yi, 1998, "Vertical specialization and the changing nature of world trade," Economic Policy Review, Federal Reserve Bank of New York.

27


Ishii, J., Yi, K.-M., 1997. “The Growth of World Trade”. Federal Reserve Bank of New York Research Paper No. 9718, May Johnson, Robert C. and Guillermo Noguera, 2012a "Accounting for Intermediates: Production Sharing and Trade in Value Added" Journal of International Economics, 86: 224236 Johnson, Robert C. and Guillermo Noguera, 2012b "Fragmentation and Trade in Value Added over Four Decades" NBER Working paper 18186 Jones, R. W. and H. Kierzkowski, 1990, “The Role of Services in Production and International Trade: A Theoretical Framework,” in Ronald Jones and Anne Krueger, eds., The Political Economy of International Trade, Basil Blackwell, Oxford. Karkkainen, A., 2008. EU-15 Foreign Direct Investment in the new member states, Eurostat, Statistics in Focus, 71. Kimura, Fukunari, 2006. “International Production and Distribution Networks in East Asia: Eighteen Facts, Mechanics, and Policy Implications,” Asian Economic Policy Review 1. Kimura F., Ando M. 2005. “Two-dimensional fragmentation in East Asia: Conceptual framework and empirics.” International Review of Economics and Finance (special issue on “Outsourcing and Fragmentation: Blessing or Threat” edited by Henryk Kierzkowski) 14. Koopman, Robert, Zhi Wang and Shang-Ji Wei, 2008 “How Much of Chinese Exports is Really Made in China? Assessing Domestic Value-Added When Processing Trade is Pervasive" NBER Working Paper No. 14109 Koopman, Robert, William Powers, Zhi Wang and Shang-Ji Wei, 2010 “Give Credit Where Credit is Due: Tracing Value Added in Global Production Chains” NBER Working Paper 16426 Krugman, Paul, 1995, “Growing World Trade: Causes and Consequences” Brooking Paper on Economic Activity, 1 Leamer, Edward E., 1996, “In Search of Stolper-Samuelson Effects on U.S. Wages,” NBER Working Paper no. 5427 Markusen, J.R., 1984. “Multinationals, Multi–Plant Economies, and the Gains from Trade,” Journal of International Economics, 16(3-4): 205–26 Markusen, J.R., 1997, "Trade versus Investment Liberalization," NBER working paper 6231, National Bureau of Economic Research Miroudot, Sebastien, and Alexandros Ragousssis, 2009, “Vertical Trade, Trade Costs and FDI”OECD Trade Policy Working Paper No. 89

28


Moreira Mesquita, Mauricio, Christian Volpe and Juan Blyde. 2008 Unclogging the Arteries: The Impact of Transport Costs on Latin American and Caribbean Trade, Special Report on Integration and Trade. Washington, DC, United States: Inter-American Development Bank. Tempest, R., 1996, “Barbie and the World Economy” Los Angeles Times, September 22. UNCTAD (2004) World Investment Report, The Shift Towards Services. United Nations. WTO, IDE-Jetro, 2011 Trade Patterns and Global Value Chains in East Asia: From Trade in Goods to Trade in Tasks. WTO and IDE-Jetro Report. Yeats, Alexander (2001) “Just How Big is Global Production Sharing” in Sven W. Arndt and Henryk Kierzkowski (Eds), Fragmentation: New Production Patterns in the World Economy, Oxford: Oxford University Press

29


Appendix

Figure1: Foreign value-added (as a share of gross exports)

40

Percentage

30

20

10

0

EU-27

Asia-Pacific

North America

LAC

Figure 2: Foreign value-added and returned domestic value-added (as a share of gross exports)

40

Percentage

30

20

10

0

EU-27

Asia-Pacific Foreign value-added

North America

LAC

Returned domestic value-added

30


Figure 3: GVC position 2

1.5

1

.5

0

LAC

North America

Asia-Pacific

EU-27

Figure 4: Regional contribution to foreign value-added 100

Percentage

80 60 40 20 0

EU-27

Asia-Pacific EU-27 North America Other

31

North America Asia-Pacific LAC

LAC


Table 1: Baseline estimation Regressor

(1)

(2)

Contiguity

0.9059*** (0.0443)

0.8989*** (0.0442)

Common language

0.2355***

0.2334***

(0.0249)

(0.0248)

Colonial ties

0.2569*** (0.0531)

0.2652*** (0.0529)

Distance

-0.7088***

-0.6667***

(0.0109)

(0.0121)

PTA

0.1475*** (0.0186)

Importing country fixed effect

yes

yes

Exporting country fixed effect Observations

yes 11740

yes 11740

0.91

0.91

2

R

*** ; ** ; * significant at the 1%, 5% and 10% level respectively

32


Table 2: The role of diagonal cumulation Regressor

(1)

(2)

(3)

(4)

Contiguity

0.8620*** (0.1092)

0.8454*** (0.1094)

0.8295*** (0.1096)

0.8295*** (0.1096)

Common language

Colonial ties Distance

0.0331

0.0245

0.0316

0.0316

(0.0639)

(0.0640)

(0.0640)

(0.0640)

0.3439* (0.1807)

0.3657** (0.1808)

0.3668** (0.1807)

0.3668** (0.1807)

-0.6488***

-0.6448***

-0.6290***

-0.6290***

(0.0426)

(0.0426)

(0.0435)

(0.0435)

0.1116** (0.0560)

0.0977* (0.0565)

PTA CUM

0.2079* (0.0186)

PTA - No cumulation

0.0977* (0.0565)

PTA - Cumulation

0.3056** (0.1222)

Importing country fixed effect Exporting country fixed effect Observations 2 R

yes yes

yes yes

yes yes

yes yes

1480 0.93

1480 0.93

1480 0.93

1480 0.93

*** ; ** ; * significant at the 1%, 5% and 10% level respectively

33


E15 GVCs Background Paper